银行卡识别与抽取（ADP）

• Outcome: Review
• Latest audit: May 28, 2026 · 8:21 AM UTC
• Version: 1.0.0

Overview

This appears to be a legitimate cloud bank-card extraction skill, but it needs Review because it handles sensitive card data and its bundled docs expose broader ADP document-processing and admin commands than the narrow card-recognition purpose suggests.

Install only if you are comfortable sending bank-card images and extracted fields to Laiye ADP cloud services under your API key. Prefer npm or verified release downloads over pipe-to-shell installers, restrict agents to the OOTB bank-card app and explicit user-provided inputs, and do not allow arbitrary folders, unrelated documents, or custom-app delete/update commands unless you intentionally want broader ADP administration.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (9)

Description-Behavior Mismatch

High

• Confidence: 94% confidence
• Finding: The README describes a broad, general-purpose ADP document-processing CLI rather than a narrowly scoped credit-card recognition skill. This scope mismatch is dangerous because agents or users may invoke capabilities far beyond the declared purpose, including arbitrary document parsing and extraction workflows, increasing the attack surface and enabling unreviewed data handling paths.

Context-Inappropriate Capability

Medium

• Confidence: 89% confidence
• Finding: Documenting custom application creation, update, and deletion introduces privileged lifecycle-management capabilities unrelated to the stated bank-card extraction function. In an agent skill context, this can let an agent alter extraction behavior or access broader platform features than intended, creating opportunities for privilege expansion or misuse.

Context-Inappropriate Capability

Medium

• Confidence: 91% confidence
• Finding: The README exposes generic remote-URL processing and broad document workflows, not just bank-card recognition. This makes the skill more dangerous in context because a user selecting a narrowly named financial-data skill could unknowingly grant an agent the ability to fetch and process arbitrary remote content or unrelated documents.

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The README describes a broad, general-purpose ADP CLI with parsing, extraction, app management, and workflow automation features that materially exceed the declared skill purpose of narrow credit-card front-side extraction. This capability mismatch can mislead users and agents into granting broader trust, invoking unintended operations, or exposing additional data and attack surface beyond the advertised function.

Description-Behavior Mismatch

Medium

• Confidence: 95% confidence
• Finding: The README advertises remote-URL processing and cloud-backed extraction despite the skill being presented as a narrow card-recognition capability. In this context, undisclosed network processing is security-relevant because sensitive financial images may be transmitted to external services or fetched from attacker-controlled URLs without users appreciating that behavior.

Context-Inappropriate Capability

Medium

• Confidence: 91% confidence
• Finding: Application-management commands such as creating, updating, and deleting custom apps are unnecessary for a fixed-purpose bank-card extraction skill and violate least privilege. Exposing these broader control-plane features increases the chance of misuse, unintended reconfiguration, or abuse of the surrounding ADP platform through a skill that users would reasonably expect to be single-purpose.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The README instructs users to submit local files and remote URLs to a cloud ADP service but does not clearly warn about privacy, confidentiality, or regulatory implications. This is especially sensitive in a credit-card context, where uploaded documents may contain PANs and personal financial data, creating risk of unintended external transmission and compliance violations.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The README recommends one-line shell and PowerShell installation commands that fetch and execute remote scripts directly, without warnings or integrity verification guidance. This pattern is dangerous because any compromise of the source repository, CDN path, transport assumptions, or user environment can lead to immediate arbitrary code execution on the host.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The skill instructs users to upload bank card images and extract sensitive payment-card data through a third-party remote ADP service, but it does not present an upfront privacy/security warning or data-handling notice before users are asked to transmit that data. In this context, the missing warning is material because bank card images and extracted fields are highly sensitive financial data, and users should be clearly informed about remote processing, retention, and compliance implications before use.

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.